From Proto-Southwestern Tai to Modern Lanna Tai: Implications From the 16th-Century Phonology

This study established the relative chronology of sound changes in the Lanna Tai language from Proto-Southwestern Tai to Modern Lanna Tai dialects. The 16th-Century Lanna Tai phonology, as documented in the Sino-Lanna Tai Manual of Translation , was compared to Proto-Southwestern Tai and Modern Lanna Tai dialects to differentiate pre-16th century and post-16th century changes. Based on this chronology, Lanna Tai can be divided into three stages: Old Lanna Tai featuring the loss of uvulars and changes in consonants; Early Modern Lanna Tai exhibiting tone split, loss of voicing contrast and vowel acquisitions; and Modern Lanna Tai dialects showing changes in consonants, acquisition of additional vowels and diphthong monophthongization in some dialects. The study focused on several sound changes, providing valuable insights into the evolution of the Lanna Tai language


Introduction
Due to its conservative orthography, existing Lanna Tai manuscripts have not been studied historically.In detail, the orthography retains a three-way tone MANUSYA: Journal of Humanities 26 (2023) 1-33 contrast and two classes of consonant graphemes reflecting original phonation types, suggesting that the writing system was adopted when there were three tones and a voicing contrast instead of the modern six tones and a lack of voicing contrast.However, the unchanged orthography makes it impossible to determine both when the three original tones split into two as well as when the voicing contrast disappeared.Furthermore, the tentative periodization of Lanna Tai by Rungreuangsri (1991) is based on writing media types rather than sound system development.Thus, our knowledge is limited to Proto-Southwestern Tai and Modern Lanna Tai, with little information about the intermediate stages.To establish a clearer developmental path, projecting an intermediate stage and identifying transitional changes will be beneficial.Continuing from Tangsiriwattanakul (2021), this study aims to establish a more precise trajectory from Proto-Southwestern Tai to Modern Lanna Tai.I will utilize evidence from the 16th-Century Lanna Tai period to demonstrate changes between Proto-Southwestern Tai and the 16th century, as well as between the 16th century and Modern Lanna Tai.Initially, I will describe the sound systems of the two known stages: Proto-Southwestern Tai and Modern Lanna Tai.By comparing these stages, I will demonstrate a detailed list of changes.This list will be compared with 16th century data to identify the preand post-16th century changes.Such an analysis will enable us to establish a relative chronology of the sound changes that shaped Modern Lanna Tai dialects over the centuries.Based on this relative chronology, I will also propose a tentative periodization of the Lanna Tai language, categorized into Old Lanna Tai (pre-16th century), Early Modern Lanna Tai (around the 16th century) and Modern Lanna Tai dialects.

Background
To comprehend the gradual sound changes from Proto-Southwestern Tai to Modern Lanna Tai, it is valuable to summarize the sound systems of both reference points: Proto-Southwestern Tai (starting point) and Modern Lanna Tai (end point).By comparing these two known stages, we can derive a list of sound changes, which will serve as the foundation for the subsequent section.

Methodology
Toward the objective of this study, the following procedures were undertaken.Firstly, the transformations that occurred from Proto-Southwestern Tai to Modern Lanna Tai were investigated by comparing their respective sound systems, as discussed earlier.This comparison yielded a comprehensive list of changes from Proto-Southwestern Tai to Modern Lanna Tai.Secondly, following the approach of Tangsiriwattanakul (2021), the vocabulary recorded in the manuscript was compared to its Proto-Southwestern Tai etyma to identify the changes from Proto-Southwestern Tai to 16th-Century Lanna Tai.This analysis differentiates the changes predating the 16th century from the changes after the 16th century.Pre-16th century changes were those present in the manuscript, while post-16th century changes were those absent from the manuscript.Refer to Table 7 for examples.
Comparing Proto-Southwestern Tai and Modern Lanna Tai reveals that *ʰr and *h in Proto-Southwestern Tai became/h/ in Modern Lanna Tai.Similarly, *ˀbl and *ˀd in Proto-Southwestern Tai both became/d/ in Modern Lanna Tai.If these changes were already completed before the 16th century, their Chinese transcriptions would be expected to be the same.However, the transcription of *ʰr differs from that of *h, while *ˀbl has the same transcription as *ˀd, as shown in Table 7.This indicates that the change *ˀbl > *ˀd predated the 16th century, whereas the change *ʰr >/h/ post-dated the 16th century.This approach allows for deriving the relative chronology of most changes.With such relative chronology established, I propose a tentative periodization of the Lanna Tai language into Old Lanna Tai (pre-16th century), Early Modern Lanna Tai (16th century) and Modern Lanna Tai dialects (present), based solely on phonological development.9 highlights the newly acquired vowels in this vowel expansion from Proto-Southwestern Tai in underlined bold print.
As mentioned previously, the six tones in Modern Lanna Tai emerged through the division of the original three tones in Proto-Southwestern Tai.The doubling of the tones was based on the laryngeal characteristics of the Proto-Southwestern Tai onset, a process comparable to the tonal development found in Sinitic, Hmong-Mien and Viet-Muong, beyond the Kra-Dai languages (Haudricourt 1954(Haudricourt , 1961)).Gedney (1972) 10 provides examples of the modern reflexes of the Proto-Southwestern Tai tones in Modern Lanna Tai.
Thus, the sound changes in (1) are defining features of Modern Lanna Tai compared to other Southwestern Tai dialects: ( Acquisition of long mid vowels/eː/,/ɤː/, and/oː/ k.Acquisition of short low front and back vowels/ɛ/ and/ɔ/ Rai dialects, ds123 = C123 for the Phrae and Nan dialects, Wimonkasem(2006) reports ds123 = A12 for the Chiang Mai dialect, so as Thianthaworn(1998, 59, 82, 103-105) for the Saraburi and Nakorn Pathom dialects.Also, Thianthaworn(1998, 59, 82, 103-105) reports ds123 = C4 for the Lopburi dialect.4 The pronunciation of the Chinese characters Tangsiriwattanakul's (2021) analysis was that of Míng-Qīng/Middle Mandarin Chinese, which was based on Jiang-Huai/Southern Mandarin dialects.During the Ming and Qing Dynasties, even after the capital was moved to Beijing, the Jiang-Huai dialect remained the official standard dialect and was known as the Nanjing dialect in Western accounts until the mid-19th century (Coblin, 2000;2002).The distinctive feature of the Jiang-Huai dialect is the fifth tone, which continued the Middle Chinese entering tone as a special tone with a final glottal stop (Norman, 2013).This feature is also found in the native rime book from the early Ming dynasty called Hongwu's Standard Rime Book (Chinese: 洪武正韻 Hóngwǔ Zhèngyùn) (Coblin, 2000;2001) Izui (1953), although the latter was never officially published.
These changes now guide the analysis of the 16th-Century Lanna Tai data.
Changes found in the manuscript indicate their completion before the 16th century, while changes absent from the manuscript suggest their occurrence after the 16th century.This approach offers insights into the timeline and relative chronology of the significant sound changes.

3.2
Lanna Tai Phonology in the 16th Century In Tangsiriwattanakul ( 2021), the 16th-Century Lanna Tai sound system was reconstructed by examining the Chinese transcription4 of Lanna Tai vocabulary in the manuscript Sino-Lanna Manual of Translation (Chinese: 八百舘譯語 Bābǎiguǎn Yìyǔ).5This manuscript serves as a glossary between Chinese and Lanna Tai, providing Lanna Tai words in Fakkham script,   palatal stops *cʰ, and a trill *r.The presence of medial *-r-in consonant clusters beginning with a stop had already changed to simple aspiration *-ʰ-.The completion or occurrence of deglottalization in preglottalized phonemes *ˀb, *ˀd, and *ˀj during that period cannot be guaranteed due to inconclusive evidence, suggesting non-contrastive realization.
The manuscript also suggests that 16th-Century Lanna Tai inherited the Proto-Southwestern Tai vowel inventory, except for the change of *aɰ to *aj.Additionally, two new vowel phonemes, *oː and *ɤː, were acquired through borrowing.However, there is no evidence for the acquisition of the vowels/ eː/,/ɛ/ and/ɔ/ found in Modern Lanna Tai.
In Tangsiriwattanakul (2021), it was demonstrated based on the correspondence between the hypothetical tones in the manuscript and the Middle Mandarin Chinese tones that the tonal inventory of 16th-Century Lanna Tai resembled that of Modern Lanna Tai.However, there is no evidence for the split of tones in dead syllables of non-voiced initials (tone D123) based on vowel length, as observed in modern dialects.This is summarized in Table 13.

3.3
Implications From Tangsiriwattanakul (2021) for 16th-Century Lanna Tai This section examines 16th-Century Lanna Tai in light of the sound changes that occurred during the transition from Proto-Southwestern Tai to Modern Lanna Tai, discussing the sound changes observed in both the pre-16th century and post-16th century periods based on the manuscript data.

3.3.1
Sound Changes Before the 16th Century In this section, the changes listed in (2) are discussed: ( Split of D4 > dl4 vs ds4 according to vowel length 16th-Century Lanna Tai most certainly underwent the devoicing of voiced obstruents, as syllables starting with voiced obstruents in Proto-Southwestern Tai were not differentiated from their voiceless counterparts in the 16th-century manuscript.This suggests the complete collapse of voicing contrast and implies the preceding tone split.Following the pattern observed in most other Southwestern Tai dialects, voiced stops simply became voiceless plain stops.Table 14 provides examples of the 16th-century reflexes of syllables with voiced initials.
The attestation of lexica in Table 15 provides further evidence for the devoicing of voiced obstruents and the voicing of voiceless sonorants.In these lexica, graphemes representing original voiced initials are used, even though they are reconstructed with original voiceless initials.For instance, the word *paːC 'aunt' is attested as <พา> bā instead of the expected <ปา> pā, and the word *ʰlajA 'flow' is attested as <ไล> lai instead of <ไหล> hlai.These instances suggest the completion of the loss of the original voicing contrast among consonants, resulting in the graphemes <ป> p and b <พ>, as well as <หล> hl and <ล> l, being consonantally identical.
Aside from the tone split and devoicing of voiced obstruents, 16th-Century Lanna Tai also exhibited the merger of *ˀbl and *ˀd.The Fakkham graphemes <บ> p and <ด> t represented two voiced stops that can be traced back to earlier *ˀb and *ˀd, respectively.For example, <บาน> pān and <ดาว> tāv cold be traced back to *ˀba:nC 'village' and *ˀdaːwA 'star' , respectively.Interestingly, words with etyma of *ˀbl were also expressed by <ด> t, for example, <เดื อน> tei̲ an Akin to Modern Lanna Tai, 16th-Century Lanna Tai also had two distinct palatal continuants: *j and *ɲ.While the 16th-century *j directly continued the Proto-Southwestern Tai *ˀj, the 16th-century *ɲ continued not only the nasal pair *ʰɲ and *ɲ but also to the original *j.This distribution suggests that the merger between *j and *ɲ into 16th century *ɲ likely predated the change from *ˀj to 16th-century *j (referred to as "bleeding order" in linguistics).If *ˀj changed to *j first (referred to as "feeding order" in linguistics), then *j and *ɲ, regardless of their origin, would both be reflected as/ɲ/ in the modern period, which is not the case.The manuscript and Modern Lanna Tai data show that *j < *ˀj remains distinct from *ɲ < *j.This also suggests that *j > *ɲ predated the voicing of voiceless sonorants.The two competing orders of operation 16th-Century Lanna Tai distinguishes between the graphemes <ข> kh and <ฃ> kh, as evidenced by their different transcriptions: <ข> kh is transcribed as Mandarin/kʰ/ while <ฃ> kh is transcribed as Mandarin/x/.However, when comparing the attested forms with the Proto-Southwestern Tai etyma, syllables with <ฃ> kh actually corresponded to three different dorsal initials in Proto-Southwestern Tai: *q, *ꭓ, and *x.On the other hand, syllables with <ข> kh corresponded to two relatively similar initials in Proto-Southwestern Tai: *kʰ and *kʰr.This reveals that Proto-Southwestern Tai *q, *ꭓ, and *x merged into 16th-Century Lanna Tai/x/, while Proto-Southwestern Tai *kʰ and *kʰr merged into 16th-Century Lanna Tai/kh/.
Likewise, the grapheme <ฅ> ɡ is transcribed as/x/, indicating that it is a devoiced form of Proto-Southwestern Tai *ɣ, whereas the grapheme <ค > ɡ is consistently transcribed as/k/, reflecting the devoiced outcome of *ɡ >/k/.However, the manuscript shows that words with the etyma of *ɢ are also written with <ฅ> ɡ and transcribed as/x/.This suggests that *ɢ > *ɣ in Lanna Tai predated *ɣ >/x/.This challenges Pittayaporn's (2009) proposal that Proto-Tai *ɢ > *ɡ in Proto-Southwestern Tai, as the reflex of *ɢ is consistently represented by *ɡ across many Southwestern Tai dialects.Based on the Lanna evidence, it is proposed that *ɢ was also present in the Proto-Southwestern Tai phonemic inventory.The merger of *ɢ with *ɡ or *ɣ occurred after the divergence of separate Southwestern Tai varieties.This distinction may serve as a criterion for sub-branching within Southwestern Tai dialects.Additionally, the presence of the word *ɢɛːp 'narrow' (etymologically related to Chinese <狹> xiá) in the manuscript as <แฅบ> ɡeep transcribed as <歇> xiē confirms Pittayaporn's (2009) reconstruction.

Feeding order Bleeding order
In Modern Lanna Tai, the vowels and final consonants have undergone minimal changes compared to the initial consonants in the Proto-Southwestern Tai phonemic inventory.Similar to Siamese, Modern Lanna Tai exhibits a merger between *aj and *aɰ, resulting in/aj/.This merger was completed by the 16th century, as evidenced by the consistent representation of syllables with *aɰ etyma as if they were *aj.For example, *ˀbaɰA 'leaf' is attested as <ไบ> pai.As a result, the Proto-Southwestern Tai finals reduced from 9 to 8, eliminating the contrast between/j/ and/ɰ/.
Unlike Siamese, but similar to Lao, Modern Lanna Tai did not undergo a merger of/ɤ/ to/ɯ/ before a velar consonant.This is evidenced by the retention of Proto-Southwestern Tai *tʰɤŋA 'arrive' and *pʰɤŋC 'bee' as/tʰɤŋ1/ and/pʰɤŋ4/ respectively in Modern Lanna Tai.The 16th-Century Lanna Tai manuscript

Proto-Southwestern
Tai Pre-16 century #1 pre-tone split Pre-16 century #2 post-tone split 16 th centurypresent also confirms this pattern.In summary, the vowel inventory changes from 14 in Proto-Southwestern Tai to 15 in 16th-Century Lanna Tai, as shown in Table 18, with the newly acquired vowel phonemes highlighted in bold and underlined.
As was already mentioned, the 16th-Century Lanna Tai exhibits a tone split and merger pattern that closely resembles Modern Lanna Tai.This indicates that the tone split occurred prior to the 16th century.Additionally, the 16th-Century Lanna Tai only shows tonal contrast in dead syllables with long and short vowels, as they are transcribed differently.This suggests that the split of tones on dead syllables with initial voiced consonants (D4 > dl4 vs ds4) predated the 16th-century.Figure 4 illustrates the change in tonal inventory from Proto-Southwestern Tai to 16th-Century Lanna Tai.

4.3.2
Sound Changes After the 16th Century This section discusses the changes listed in (3): (3) Sound changes after the 16th century: a.As previously mentioned, Proto-Southwestern Tai *q, *ꭓ, and *x merged into the 16th-Century Lanna Tai *x as represented by the grapheme <ฃ> kh, whereas the Proto-Southwestern Tai *kh and *khr coalesced into the 16th-Century Lanna Tai *kʰ, as represented by the grapheme <ข> kh.In addition, the phoneme *x as represented by the grapheme <ฅ> ɡ can be traced back to Proto-Southwestern Tai *ɣ & *ɢ as well.The contrast between *kʰ vs *x in the 16th century suggests that the change form *x > *kʰ actually post-dated the 16th century.At this point, became possible to schematize the path of change of the Proto-Southwestern Tai dorsal consonants as shown in Table 19.
In the manuscript, the grapheme <ฉ> ch appears five times, and each instance is transcribed as Mandarin/ʈʂʰ/.This indicates the presence of the *cʰ in the 16th-Century Lanna Tai.Although Modern Lanna Tai no longer has initial/cʰ/, and words in Siamese with initial/cʰ/ are now pronounced with initial/s/ in Modern Lanna Tai, there was a clear distinction between the two sounds in the 16th century, since the 16th-Century *s transcribed as Mandarin/s/, never overlapped with the *cʰ as represented by <ฉ> ch.While *cʰ was not originally present in Proto-Tai or Proto-Southwestern Tai, it was likely acquired through contact with Chinese and Austro-Asiatic languages.The five instances of lexica with initial <ฉ> ch cannot be traced back to a Tai etymon.However, the presence of this phoneme was likely marginal and short-lived, as it merged with/s/ in most Modern Southwestern Tai varieties.Only Siamese and Southern Thai, which have had prolonged contact with Austro-Asiatic languages, have  retained this phoneme in their systems.The merger of the newly acquired *cʰ with the existing *s must be relatively recent in Lanna Tai, as *cʰ still existed as a distinct phoneme in the 16th century.
In Modern Lanna Tai, the reflex of Proto-Southwestern Tai voiceless *ʰr is identical to Siamese, represented as/h/.However, the manuscript reveals an intriguing and unexpected development.The consistent transcription of <หฺ ร > hr and <ร> r as Mandarin/l/ indicates the presence of *r.Unlike Siamese, where *ʰr changed to/h/ as early as the late 13th century, *ʰr likely remained as *ʰr until it merged with the original *r, hence its transcription as/l/.This data suggests that while the voicing of voiceless sonorants and devoicing of voiced obstruents predated the 16th century, the merger of *r to *h actually post-dated the 16th century, when the manuscript was composed.
The phonetic nature of the 16th-Century *j, which evolved from Proto-Southwestern Tai *ˀj, as either a plain approximant [j] or a pre-glottalized [ˀj], remains unknown.Similarly, it is unclear if *d and *b in the 16th century were still pre-glottalized or had de-glottalized to simple voiced sounds.The de-glottalization of *ˀj > *j does not necessarily coincide with the de-glottalization of pre-glottalized stops *ˀb > *b and *ˀd > *d, as some Modern Tai dialects retain the pre-glottalization of *ˀb and *ˀd, while virtually none preserve the pre-glottalization of *ˀj.Relatively though, the de-glottalization of *ˀb and *ˀd must have post-dated the devoicing of voiced obstruents for they evaded merging with the original voiced stops.If the de-glottalization predated the devoicing, they would have become devoiced like the original voiced stops and would now be indistinguishable from one another, as shown in Table 22.Similarly, the merger of *ˀbl with *ˀd must have preceded the de-glottalization of *ˀb and *ˀd, as it yielded the same result as *ˀd. Figure 5 illustrates the sequence of phonological changes involving pre-glottalized and plain voiced stops in Proto-Southwestern Tai.
The 16th-Century Lanna Tai had only acquired long *oː and *ɤː, whereas the Modern Lanna Tai has a more complete and symmetrical vowel inventory.This suggests that the acquisition of long/eː/, short/ɛ/, and short/ɔ/ post-dated the 16th century.Table 24 provides an overview of the vowel system in Modern Lanna Tai, highlighting the vowels that were acquired after the 16th century with underlined bold print.Dialectologically, Modern Lanna Tai dialects exhibit the merger of *B tones and tones in long-dead syllables (dl tones) but they differ in how the tones in short dead syllables (ds tones) merged with tones in non-checked syllables.The split of tones on dead syllables based on the laryngeal settings of the initial consonant (i.e., tone *D > D123 & D4) occurred before the loss of the original voicing contrast.While Modern Lanna Tai dialects show the split based on vowel length for both originally voiced (D4 > dl4 vs ds4) and non-voiced consonants (D123 > dl123 vs ds123), the 16th-Century Lanna Tai shows such a split for originally voiced consonants only.This suggests that the split based on vowel length for originally non-voiced consonants postdared the 16th century.Figure 6 illustrates the tonal changes from the 16th century to Modern Lanna Tai.
Furthermore, while modern Lanna Tai dialects exhibit a shared tone split and merger pattern, the pitch height and contour of the 16th-Century Lanna Tai tones evolved into distinct outcomes in each Modern Lanna Tai dialect.However, this study does not delve into the specific changes in the phonetic realization of each tone from Proto-Southwestern Tai to Modern Lanna Tai dialects.

5.1
Old Lanna Tai I propose the period of Old Lanna Tai as the oldest stage of Lanna Tai diverging from Proto-Southwestern Tai.Though its existence predates the 16th century, the exact time of divergence from Proto-Southwestern Tai is still unknown.The distinguishing features separating Old Lanna Tai from its proto-language include those in ( 4 25, while the vowels and tones remained stable.

5.2
Early Modern Lanna Tai I propose that Early Modern Lanna Tai can be conceived as a transitional phase between the Old Lanna Tai and the Modern Lanna Tai dialects.During this period, significant changes occurred, including the split in tones and the loss of the original voicing contrast.The voiced-voiceless minimal pairs merged, while the existing tones split based on the variation in voicing of the initial consonants.The number of tone phonemes doubled from 3 to 6, while the number of consonant phonemes decreased from 38 to 23.The timing of the de-glottalization of pre-glottalized phonemes *ˀb, *ˀd, and *ˀj is still unknown.In terms of vowels, Early Modern Lanna Tai acquired the long mid non-front vowels *oː and *ɤː and the distinction between *aj and *aɰ was lost in favor of a single/aj/.The split of Tone D4 according to vowel length can be confirmed based on the different tonal transcriptions in the manuscript.Taking these changes into account, the inventories presented in Table 25 transformed into the ones shown in Table 26.Additionally, Table 27 illustrates the tone inventory and demonstrates the patterns of tone split and merger.

Conclusion
This study enhances our understanding of the phonological history of the Lanna Tai language.By comparing earlier research with the known endpoints in the phonological history of Lanna Tai, this study provides a detailed analysis of the development from Proto-Southwestern Tai to Modern Lanna Tai dialects.It also uncovers a relative chronology of the sound changes that characterize Modern Lanna Tai dialects and based on the relative chronology of these changes, tentatively periodizes Lanna Tai's phonological history into three periods.This study follows in the footsteps of previous studies that utilize philological evidence to illustrate language development.Further replication of such studies for other Tai languages with extensive written evidence is strongly encouraged.
figure 4 Tonal Changes from Proto-Southwestern Tai the 16th-Century Lanna Tai

table 9
Modern Lanna Tai vowel inventory as symmetricized from Proto-Southwestern Tai tangsiriwattanakul 10.1163/26659077-26010011 | MANUSYA: Journal of Humanities (2023) 1-33 Downloaded from Brill.com 03/30/2024 09:28:44AM via Open Access.This is an open access article distributed under the terms of the CC BY 4.0 license.https://creativecommons.org/licenses/by/4.0/ 3 Whereas Akharawatthanakun(2012, 178) reports ds123 = A34 for the Lampang and Chiang and European manuals for "Mandarin Chinese" like Francisco Varo's Arte de la Lengua Mandarina (1703).For the purpose of this study, the Romanized transcription of the Chinese characters is provided in Modern Standard Mandarin Pinyin.5 The version examined in Tangsiriwattanakul (2021) is the one preserved at the National Library of China, available online at https://archive.org/details/02076757.cn.This version is identical to the digitized Berlin version found at the Staatsbibliothek zu Berlin, accessible at https://digital.staatsbibliothek-berlin.de/werkansicht?ppn = ppn334615730 x&physid = phys_1509&d mdid = dmdlog_0008&view = overview-toc (listed as 8.Pa-pai, Suppl Vol.15; Texte Vol.21).Both versions match the Toyo Bunko version mentioned in

table 10 Modern
Lanna Tai reflexes of Proto-Southwestern Tai tones

table 14
Reflex of original voiced stops in the 16th century are compared in Table16, while Figure2illustrates the chain of phonological changes in Proto-Southwestern Tai palatal continuants.

table 16
Order of operations for the palatal mergers

table 18
Simple vowel inventory of the 16th-Century Lanna Tai

table 19
Proposed timeline of changes in dorsal obstruents Downloaded from Brill.com 03/30/2024 09:28:44AM via Open Access.This is an open access article distributed under the terms of the CC BY 4.0 license.https://creativecommons.org/licenses/by/4.0/

table 23 From
Proto-Southwestern Tai consonants to Modern Lanna Tai consonants

table 29 Modern
Lanna Tai tones